ABSTRACT
The legend: A cross-level direct effect/path from some higher-order situational attribute to some lower-order individual variable may be statistically estimated. The kernel of truth: Conceptually, yes, an effect from the higher-order attribute to some lower-order effect may be hypothesized. The myth: The observed parameter estimate taught to researchers as the one representing the cross-level direct effect is the direct conceptual effect from the situational higher-order attribute to the lower-order individual effect. The follow-up: If the within-group mean is truly representative of all individuals in a group, then extremely strong within-group agreement should be present. Only under those conditions will the “direct-effect” parameter estimate represent the theoretical cross-level direct effect.
